Re: Mapping Behavioral Traits

From: Daniel Collico Savio <CollicoD_at_advance.com.ar>
Date: Mon 15 Jan 2001 14:04:43 -0500

MARTYN EMERY <martyn_emery_at_hotmail.com> WROTE:
> It's day 5 of my new health kick, and those billion
> brain cells that I 'lost'due to seasonal alcohol abuse
> are arriving back en-mass, hence the blue sky nature
> of this posting.

Good for you, Martyn. Your alcohol level in blood
should be following right now a descending Gaussian
curve, which would probably reach zero level in a week
or so. Cheers, indeed.

> I would be interested to know how many of the list are
> currently involved in mapping the different behavioral
> traits of their clients both offline and online.

My experience is related to online clients (more
generally, users). Two years ago the TopMail -a free
webmail- was launched in Argentina. In our company we
had neither DataMining soft tools nor human resources
to identify TopMail users. Don´t want to sound
naive, but there was a new land to be explored: a huge
database of thousands of rows and twenty or thirty
columns (demographic and shopping habits provided by
users in the sign up).

I divided the data in "pieces" of sixty thousand rows,
the largest amount an Excel or Access can deal with.
Then I followed the basic guidelines of the Forrester
Technographics segmentation process, which states that
several "clusters" of users can be obtained by
filtering the gross data in terms
of...

a) Users primary motivation (career, family, entertainment)
b) Income (High, low)
c) Attitude towards technology (Optimists, Pessimists)

The original Forrester article is no longer in the web
site, but some ideas can still be found at
http://www.forrester.com/Products/Techno/0,3833,0,FF.html

Of course, this is an arbitrary methodology -as any-.
For instances, the limit between "high" and "low"
income depends on each country. The main idea here is
that you are not supossed to pay huge budgets to
consultant firms, just to know your clients. But if
your company has the budget... well, it´s OK. Maybe
it's all about the data volume and the time involved in
the research.

Best regards, Martyn. Go ahead with your descending
gaussian -and that blue sky nature-.

Daniel Collico Savio
collicod_at_advance.com.ar
Buenos Aires, Argentina
Telefonica





Received on Mon Jan 15 2001 - 13:04:43 CST